Luminescence dating: laboratory procedures and protocols

In the last 30 years, from 1967 to 1997, the use of luminescence signals from naturally occurring minerals has gone though a major metamorphosis, from thermoluminescence (TL) dating of pottery to optically stimulated luminescence (OSL) dating of sediments. Laboratory procedures for dating sediments have been adapted from those for pottery and new procedures have been developed as the need arises.

The majority of sediment dating applications are carried out on quartz and potassium-rich feldspars and the general characteristics of the TL and OSL signals from these minerals are reviewed. For sediments some new problems were encountered, with some grains perhaps not being completely bleached at deposition. For OSL signals there is no simple procedure for the selection of a thermally stable signal, as there had been in the case of pottery.

Many different laboratory protocols have been developed as our understanding of the fundamental behaviour of luminescence signals from quartz and feldspar has improved. These protocols are explained and discussed, giving the advantages and disadvantages of each procedure as applied to different types of sediment.

This review is presented as a guide to the selection of the most appropriate procedure for a particular dating application.     

Paper I: The use of measurement-time dependent single-aliquot equivalent-dose estimates from quartz in the identification of incomplete signal resetting

The identification of incomplete signal re-setting of optically stimulated luminescence signals in sedimentary quartz is a vitally important step in the continued improvement of optical dating. It is shown that narrow spectrum (blue-green) laboratory partial bleaching of aliquots of natural sedimentary quartz has a significant effect on equivalent dose (measured using a single aliquot procedure) calculated as a function of measurement time (D_e(t)). The blue-green stimulation spectrum mimics that found underwater and the results suggest that incomplete re-setting of waterlain sediments may be possible using the D_e(t) method.     

Luminescence property of volcanic quartz and the use of red isothermal TL for dating tephras

An optically stimulated luminescence (OSL) age obtained from a Japanese tephra using quartz phenocrysts severely underestimated the known age. The characteristics of the OSL signals were investigated in order to understand the cause of the underestimation; the main OSL component of volcanic quartz has a thermodynamic lifetime of about 1700 years at room temperature, and it also seems to fade anomalously (i.e. athermally). Measurement of conventional red thermoluminescence (RTL) using a Ga–As photomultiplier tube was difficult due to the presence of a strong thermal background, although RTL gave an age consistent with the independent age. Furthermore, red isothermal TL (RITL) at 380 °C allowed the RTL signal to be separated from to the thermal background, and RITL ages of three volcanic quartz samples show good agreement with independent ages.     

Investigation on non-optically bleachable components of ESR aluminium signal in quartz

Electron Spin Resonance (ESR) can be used as a method to estimate the age of sediment deposition using the paramagnetic centre related to aluminium impurities in quartz. This so-called Al-centre can be partially optically bleached and its signal intensity decreases in relation to time exposure to solar light, until it reaches a plateau value corresponding to a residual signal. This signal can be attributed to “Deep Aluminium Traps” (DAT) which cannot be reset by an exposure to sunlight. In this study, we have investigated the behaviour of the DAT signal in samples from different origins and ages. The intensity of the DAT signal has been isolated from the total aluminium signal by the exposure of different quartz samples to simulated solar light. We observed that the DAT intensities were sample dependant and therefore it should be determined for each sample. Moreover, DAT intensities of Pleistocene volcanic quartz increase with gamma laboratory irradiation, whereas DAT intensities of sedimentary quartz do not vary with added artificial doses. This suggests that DAT in quartz extracted from sediments must be inherited from the primary source of the quartz, and were saturated at the time of sedimentation. We thereby validate the ESR dating of quartz sediment protocol used so far.     

Behavioural studies of stimulated luminescence from feldspars

Since the work of Mejdahl in the mid-1980s feldspars have been used widely for palaeodosimetry. They have proved particularly popular because of the technical ease with which an optically stimulated luminescence signal can be obtained, and the potential for their use in dating over a wider time range than quartz. However, while the use of infrared stimulated lminescence has been a great success, the realisation of the potential for dating over a wide time range has been hampered by a series of behavioural problems, in particular concerned with the stability of the luminescence signal. This paper surveys the current state of knowledge of a range of aspects of the stimulated luminescence behaviour of feldspar. Particular attention is paid to the thermal and optical properties of the thermoluminescence (TL) and optically stimulated luminescence (OSL) signals from feldspars since these are of greatest relevance to their use as palaeodosimeters. Results obtained from well-characterised museum specimens demonstrate the variety of responses that are seen from different types of feldspar, particularly in their TL behaviour. Less variation is seen in their OSL response. In contrast, the response of potassium-rich feldspars separated from Quaternary sediments tends to be very consistent when many grains are analysed simultaneously, suggesting that averaging of the signal simplifies the behaviour of the system.     

Fluvio-mechanical resetting of the Al and Ti centres in quartz

ESR dating of fluvial terraces is usually based on the assumption of sunlight resetting of the Al centres in quartz. Very long bleaching times (months of sunlight exposure) are required to reset this centre to a stable, non-zero ESR intensity. Considering that highly light sensitive optically stimulated luminescence (OSL) analyses of fluvial samples often indicate partial bleaching, ESR dating should be extremely difficult. Two samples were analysed, a fluvial sediment from a Chinese river terrace and an Australian granite. Using a Höhnle sunlight simulator, the Ti centres of both samples were completely bleached within 1 to 65 h, while the Al centre required extreme bleaching times to reach stable residual intensities (∼1000 h). As an alternative to light resetting, tumbler experiments were carried out to simulate fluvial action. The samples were treated under OSL sample preparation conditions. After relatively short tumbling times, corresponding to a few km of fluvial transport, both Al- and Ti-centres were partially reset. In a granite sample, the tumbling reduced the intensity of the Al centre more effectively than light exposure.     

Quartz luminescence response to a mixed alpha-beta field: Investigations on Romanian loess

Previous SAR-OSL dating studies using quartz extracted from Romanian and Serbian loess samples report SAR-OSL dose–response curves on fine grained (4–11 μm) quartz that grow to much higher doses compared to those of coarse-grained (63–90, 90–125, 125–180 μm) quartz. Furthermore, quartz SAR-OSL laboratory dose response curves do not reflect the growth of the OSL signal in nature. A main difference in coarse- and fine-grained quartz dating lies in the alpha irradiation history, but the effect of mixed alpha-beta fields has so far received little attention. In the present study we investigate whether the alpha dose experienced by fine grains over geological cycles of irradiation and bleaching may have an effect on the saturation characteristics of the laboratory dose response. By applying time resolved optically stimulated luminescence we confirm that the OSL signals induced in quartz by alpha and beta radiation follow the same recombination path. We also show that a mixed alpha-beta dose response reproduces the beta dose response only up to about 800 Gy. Assuming an a-value of 0.04 we have shown that laboratory alpha and beta dose response curves overlap up to effective alpha doses of ∼50 Gy. Based on these results, we conclude that exposure of fine grains to alpha radiation during burial and transport cycles prior to deposition, as well exposure to the mixed radiation field experienced during burial are not responsible for the age discrepancies previously reported on fine and coarse grained quartz extracted from Romanian and Serbian loess.     

Sediment dating by luminescence: a review

In the present article we review applications of luminescence methods for the determination of the age of sediments for Quaternary Earth science. “Sediment” is taken to include any mineral particles transported by wind, water and/or ice and subsequently deposited. Methodology is not discussed per se but appears in context. The emphasis is on recent work and is illustrative rather than exhaustive. Applications in archaeology are mentioned only briefly. Successful dating requires that the luminescence signal has been zeroed by sunlight at the time the sediment was laid down. Dune sands and loess transported in full daylight usually satisfy this condition and have been dated by both traditional thermoluminescence and by optical methods. Optical dating is advantageous for less well-bleached sediments, such as are found in colder climates and those laid down by fluvial and glacial processes, although the prospects for dating the latter are shown to be poor. Among less common contexts are volcanic eruptions, earthquake faults and tsunami deposits. The dating of very young sediments, less than 2 ka, is demonstrated and the factors limiting the dating of very old sediments, 500 ka to 1 Ma, are discussed. Validation of any dating method by comparison with other methods is necessary; luminescence ages have been successfully compared with ages obtained by, for example, ¹⁴C, U–Th series and δ¹⁸O isotope ratios. The most difficult problems encountered in dating quartz, feldspars and undifferentiated fine grain mixtures include: anomalous fading (which leads to underestimates of age), incomplete zeroing (which gives overestimates) and sample inhomogeneity. Methods devised to minimise these effects include preheating regimes, selection of observing wavelengths and of stimulating wavelengths for optical dating, the use of single aliquot and single grain measurements. The use of appropriate data analysis procedures is also important. Advice on current best practice on procedures for obtaining reliable ages is offered and, in all applications, there are suggestions as to where further research might be directed.     

Natural and laboratory TT-OSL dose response curves: Testing the lifetime of the TT-OSL signal in nature

This study compares natural and laboratory generated thermally transferred optically stimulated luminescence (TT-OSL) dose response curves (DRCs) for fine-grain quartz extracts from the Luochuan loess section in central China. Both DRCs saturate at high doses relative to the quartz OSL signal; the natural TT-OSL DRC saturates at about 2200 Gy and laboratory DRCs saturate at about 2700 Gy. However, the natural and laboratory TT-OSL DRCs deviate from one another at circa 150 Gy resulting in TT-OSL equivalent dose underestimation relative to palaeodoses expected from dose rates and independent age control. The lifetime of the TT-OSL signal at 10 °C, calculated from values of trap parameters E and s, is compared against the value for lifetime of the TT-OSL signal in nature at average burial temperature as determined from the age underestimation caused by deviation of the natural and laboratory generated DRCs. These two independent assessments of TT-OSL signal lifetime at Luochuan give similar values, suggesting that laboratory measurements of thermal stability reflect natural burial lifetimes and can potentially be used to correct TT-OSL ages for the difference between natural and laboratory dose response curves.     

Assessing the impact of pulsed-irradiation procedures on the thermally transferred OSL signal in quartz

Thermally transferred optically stimulated luminescence (TT-OSL) dating protocols have been suggested as a means of extending the age range of luminescence dating. Several studies demonstrate that TT-OSL signals increase with large radiation doses (>2000 Gy) and yet, few studies report older TT-OSL ages (>400 Gy) in agreement with independent absolute age control. In one such study, agreement with independent chronology was only achieved for the old samples by implementing a pulsed-irradiation procedure. Pulsed-irradiation is suggested to compensate for dose rate dependent competition effects by dividing the laboratory irradiation into discrete irradiation steps interspersed with heat treatments. However, every inter-step heat treatment has the potential to anneal part of the TT-OSL dating signal. This study compares constant- and pulsed-irradiation TT-OSL protocols and investigates the degree of partial thermal annealing. The results suggest that almost all of the difference in outcome between constant- and pulsed-irradiation protocols can be explained by partial annealing of the TT-OSL signal rather than by competition effects. Partial annealing distorts the laboratory dose response curve but has no impact on the natural signal, resulting in unreliable equivalent dose estimates. This means that pulsed-irradiation procedures may not be viable for TT-OSL dating measurements. Future studies implementing pulsed-irradiation procedures should carefully consider the extent to which inter-step thermal treatments partially anneal the signal.     

Use of the LM-OSL technique for the detection of partial bleaching in quartz

We present a study of the sensitivity to light (ease-of-bleaching) of the trapped charge in sedimentary quartz grains using an optically stimulated luminescence (OSL) technique in which the intensity of the stimulation light is linearly increased during the measurement period. The technique is known as linear modulation OSL (LM-OSL). In controlled laboratory conditions, this technique has been employed to study the ease-of-bleaching of the trapped charge in quartz by comparing the OSL curves of quartz aliquots which have been either: (1) fully bleached, followed by a laboratory dose of β-irradiation, or (2) partially bleached, followed by the laboratory β-dose. The ratio of the OSL signals due to the β-dose from the partly and fully bleached aliquots is illustrated to be a potential indicator of the degree of optical resetting of the OSL signal in dating material. The key parameter governing the ease-of-bleaching is the photoionization cross-section of the trap involved. The concept is also demonstrated in a model study from which very good agreement with the experimental observations has been found. Potential applications of the technique to dating are discussed.     

沉积物中石英ESR测年功率饱和效应的初步研究

介绍了几组沉积物中石英ESR功率饱和效应的基础实验情况,认为在用石英进行ESR测年时,微波功率的选择是一个重要参数,它直接影响累积剂量(AD)的求取,从而严重影响测年结果。在某些特定的地质环境下(如黄土,海洋沉积物),采用比较高的微波功率,可从沉积物石英的E＇心得到较老沉积物(大于50万年)可信的ESR年代。     

沉积物中石英ESR测年功率饱和效应的初步研究

介绍了几组沉积物中石英ESR功率饱和效应的基础实验情况,认为在用石英进行ESR测年时,微波功率的选择是一个重要参数,它直接影响累积剂量(AD)的求取,从而严重影响测年结果。在某些特定的地质环境下(如黄土,海洋沉积物),采用比较高的微波功率,可从沉积物石英的E'心得到较老沉积物(大于50万年)可信的ESR年代。     

The timing of OSL sensitivity changes in a natural quartz

The timing of 110°C thermoluminescence (TL) and optically stimulated luminescence (OSL) sensitivity changes is explored using a natural (aeolian) quartz sample from Australia that was previously found to exhibit marked dose-dependent sensitivity change. The changes occur asynchronously; the 110°C TL sensitivity changes after dosing plus pre-heating, whereas the OSL sensitivity changes (proportionately with dose) after optical bleaching. Although the magnitude and saturation characteristics of the sensitivity changes are found to be similar, their non-synchronicity negates a direct link between the phenomena. Implications of these results for dating procedures are discussed. In particular, the results indicate that a single aliquot additive dose procedure similar to that developed for coarse-grained potassium feldspar should be possible for some quartz samples.     

Electron Spin Resonance (ESR) Dating of Calcareous Fault Gouge of the Ushikubi Fault, Central Japan

The ages of fault events of active faults have been estimated using electron spin resonance (ESR) signals of siliceous gouges. This technique of ESR method is limited by obtaining only ages that are greater than tens of millennia. So this study focuses on developing a new technique of using calcareous gouges to gain an insight into the ages of latest seismogenic event within the Holocene. For the first time, signal B of the ESR method has been used to estimate the age of the Ushikubi fault from calcareous gouge. This technique proved reliable because the mean age (1.9 ka) obtained agrees with previous works on indirect age determination of latest fault events by utilizing radiocarbon dating in the study area. However, the result from the ESR technique showed an increase relative to the age of 1 ka that was obtained by the radiocarbon dating method. This disparity may be due to a high dose rate value of 50 Gy/h of artificial irradiation that was used to determine the equivalent dose (ED). Moreover, isochronal experiment revealed that the gouge did not comprise pure carbonates but consisted of a mixture of calcite and quartz grains. A younger age value would have been obtained if a lower artificial irradiation dose rate and a relatively pure carbonate fault gouge were used in the ED determination.     

OSL properties of NaCl relative to dating and dosimetry

The luminescence properties of NaCl are discussed. Attention is focussed predominantly upon the optically stimulated luminescence (OSL) signal (UV emission) of NaCl (420–560 nm stimulation). The results from experiments relevant to geological dating applications, such as signal resetting, dose response and thermal stability, are described. The behaviour of the OSL signal observed was found to be favourable for dating, with both the thermal stability and dose response suggesting a range of at least 10 ka. Signal resetting via optical exposure was found to be extremely rapid and no evidence of either thermoluminescence or OSL signals was found following re-crystallization. The potential for dating natural salt deposits, as found in various arid environments, appears to be good.     

Potential of the slow component of quartz OSL for age determination of sedimentary samples

The slow component of quartz OSL exhibits a high thermal stability, and, in many of the samples studied, a high dose saturation level (several hundreds or, even thousands, of Grays). These properties suggest that the slow component has potential as a long-range dating tool. Initial attempts have been made to obtain equivalent doses for a number of sedimentary samples. Single- and multiple-aliquot techniques were modified for use with the slow component. The results indicate that there is a good potential for sediment dating, particularly for samples of significant age. Experiments concerning the optical resetting of the slow component suggest that, given its slow optical depletion rate, dating may be restricted to aeolian sediments.     

Na-rich feldspar as a luminescence dosimeter in infrared stimulated luminescence (IRSL) dating

One of the challenges in dating rock surfaces is the choice of the luminescence mineral. Although quartz is the preferred dosimeter in sediment dating, it is often not sufficiently sensitive when extracted from solid rocks. The intensity of signals from feldspars tends to be much less dependent on geological origin and erosion history, but the dosimetry of K-rich feldspar grains extracted from rocks is complicated because the internal dose rate is very dependent on the original feldspar grain size. The in situ grain size information is lost during the crushing process used to separate the grains for measurement. This latter problem does not apply to Na-rich feldspar because of the absence of internal radioactivity.The potential application of Na-rich feldspar as a luminescence dosimeter for the IRSL dating of rock surfaces is investigated using a variety of sediment samples from different geological settings for which independent age control is available. The blue and yellow luminescence emissions are measured for IR stimulation at 50 °C (IR₅₀), and post-IR IR stimulation at 290 °C (pIRIR₂₉₀). Thermal stability experiments imply that the corresponding signals in both emissions have comparable thermal stabilities and that all signals have similar recombination kinetics and are thermally stable over geological timescales. The IR₅₀ doses measured using blue and yellow emissions are similar to or lower than quartz doses while pIRIR₂₉₀ blue doses are higher than those from yellow emission and quartz doses. The fading rates measured for the IR₅₀ signals are ∼3%/decade larger than those measured for the pIRIR₂₉₀ signals in both yellow and blue emissions. Furthermore the average fading rates of both yellow signals are ∼3%/decade higher than the corresponding fading rates of the blue signals. However, there is no detectable correlation between fading rates and the measured D_e values. The residual doses measured from the laboratory-bleached samples and a modern analogue suggest that the IR₅₀ signals in both blue and yellow emissions bleach to the same degree, as do the corresponding pIRIR₂₉₀ signals, and that there is no significant naturally-unbleachable residual dose observed using these signals. Neither anomalous fading nor incomplete bleaching explains the observed dose discrepancy between the two emissions. Eight uncorrected and fading-corrected ages are calculated for each sample based on all four signals, using the dose rate relevant to Na-rich feldspar extracts (i.e. ∼3% K). The IR₅₀ and pIRIR₂₉₀ blue ages were also calculated assuming a dose rate based on 12.5% internal K (i.e. assuming that the blue signals were mainly derived from contamination by K-rich feldspar). The latter pIRIR₂₉₀ blue ages are in agreement with the expected age control, raising the possibility that this signal originates mainly from K-rich feldspar contamination in our Na-rich fractions, and thus is not so useful in the luminescence dating of rock surfaces. On the other hand, the pIRIR₂₉₀ fading-corrected ages based on the yellow emission are consistent with the independent age controls; higher preheat and stimulation temperatures may result in more stable yellow signals from Na-rich feldspar extracts from rocks, and so reduce the size of the fading correction. We conclude that, because this signal avoids the dosimetry difficulties of K-rich feldspar extracts, it has considerable potential in the IRSL dating of rock surfaces.     

Investigating quartz optically stimulated luminescence dose–response curves at high doses

Despite the general expectation that optically stimulated luminescence (OSL) growth should be described by a simple saturating exponential function, an additional high dose component is often reported in the dose response of quartz. Although often reported as linear, it appears that this response is the early expression of a second saturating exponential. While some studies using equivalent doses that fall in this high dose region have produced ages that correlate well with independent dating, others report that it results in unreliable age determinations. Two fine grain sedimentary quartz samples that display such a response were used to investigate the origin of this additional high dose component: three experiments were conducted to examine their dose–response up to >1000 Gy. The high dose rates provided by laboratory irradiation were found not to induce a sensitivity change in the response to a subsequent test dose, with the latter not being significantly different from those generated following naturally acquired doses. The relative percentage contributions of the fast and medium OSL components remained fixed throughout the dose–response curve, suggesting that the electron traps that give rise to the initial OSL do not change with dose. An attempt was made to investigate a change in luminescence centre recombination probability by monitoring the depletion of the ‘325 °C’ thermoluminescence (TL) during the optical stimulation that would result in depletion of the OSL signal. The emissions measured through both the conventional ultraviolet (UV), and a longer wavelength violet/blue (VB) window, displayed similar relative growth with dose, although it was not possible to resolve the origin of the VB emissions. No evidence was found to indicate whether the additional component at high doses occurs naturally or is a product of laboratory treatment. However, it appears that these samples display an increased sensitivity of quartz OSL to high doses that is not recorded by the sensitivity to a subsequent test dose, and which results in a change in the sensitivity-corrected dose–response curve.     

Identifying well-bleached quartz using the different bleaching rates of quartz and feldspar luminescence signals

When dating older sedimentary deposits using quartz, there are no unambiguous methods for identifying the presence of incomplete bleaching. Current statistical analysis of dose distributions depends entirely on the assumption that incomplete bleaching and mixing are the main causes of any excess dispersion in the distribution; the only existing way to test this assumption is using independent age control. Here we suggest a new approach to this question, based on the differential bleaching rates of quartz and feldspar luminescence signals. We first present data that confirm the differences in relative bleaching rates of quartz optically stimulated luminescence (OSL) and feldspar luminescence stimulated at 50 °C by infrared light (IR₅₀) and feldspar luminescence stimulated at 290 °C by infrared light after a stimulation at 50 °C (pIRIR₂₉₀), and use recently deposited samples to determine the likely significance of the difficult-to-bleach residual feldspar signals in non-aeolian samples. For a set of mainly Late Pleistocene non-aeolian sediments, large aliquot quartz doses are then used to predict feldspar doses (based on a knowledge of the sample dose rates). The differences between observed and predicted feldspar doses as a function of the quartz dose, combined with a conservative assumption concerning the relative feldspar and quartz residual signals after natural bleaching prior to deposition, are used to identify those samples for which the quartz is very likely to be well bleached (20 out of 24). Two of these apparently well-bleached samples are then examined using single-grain quartz dose distributions; one of these is consistent with the well-bleached hypothesis, and one indicates poor bleaching or a multi-component mixture. However, independent age control makes it clear that the large aliquot data are more likely to be correct. We conclude that a comparison of quartz and feldspar doses provides a useful independent method for identifying well-bleached quartz samples, and that it is unwise to apply statistical models to dose distributions without clear evidence for the physical origins of the distributions.